摘要:

668 J. Chem. SOC. (A), 1968 Polarised Crystal Spectra of Some Dimeric Palladium(ii) and Platinum(ii) Halide Complexes By P. Day, M. J . Smith, and R. J. P. Williams, University of Oxford, Inorganic Chemistry Laboratory, South Polarised transmission spectra are reported for single crystals of bistetraethylamrnoniurn hexabromo- and hexaiodo- dipalladate(I1) and the corresponding diplatinates. The known crystal structure of the hexabromodiplatinate(l1) permits the resolution of the polarisation of the ligand-field transitions along the molecular axes and hence leads to assignments of the excited states. The relationship between the assignments for all four compounds and those for tetrachloroplatinate(1 I ) and tetrachloropalladate( 11) is discussed. There is no spectroscopic evidence for electronic interaction between the metal atoms in the dimeric anions.Parks Road, Oxford THE possibility of metal-metal bond formation between the square planar ds complex ions which form stacks in such compounds as Magnus’ Green Salt, Pt (NH,),PtCl,, and nickel dimethylglyoxime has been widely discussed.l The polarised crystal spectrum of Magnus’ Green Salt, however, was interpreted as resulting from ligand-field transitions of PtC142-, and no extra bands were observed which could be attributed to transitions to charge trans fer or metallic conduction band excited states, with the exception of a particularly weak band in the near in- frared. Work on other platinum(@ ainmines , showed that the latter was most probably an N-H overtone vibration. To interpret the crystal spectrum of Magnus’ Green Salt, reference was made to the crystal spectrum of K2PtC1,, which has received a good deal of attention from other author^.^-^ In this compound, too, the PtC1,2- groups are stacked in columns with their planes parallel, though the metal-metal distance is much greater (4.10 A) than in Magnus’ Green Salt, and no spectro- scopic evidence has ever been found to suggest that metal- metal bonding occurs.Nevertheless, the polarised spectra of crystals in which PtIIX, groups are not stacked plane to plane would be of interest. The crystal struc- ture and reported crystalline habit of [N(C2H5),I2Pt2Br6 suggest its suitability for polarised crystal spectroscopy since the compound is triclinic, with one molecule per unit cell, so that factor group splittings can be ignored, and the planes of the Pt2Br,2- groups make only a small angle to the well developed (100) face of the plate-like crystals. There is also the interesting question whether any metal-metal bonding might be detectable within each dimeric unit, where the distance between the two platinum atoms is approximately 3.4 A.In addition to the hexabromodiplatinat e (11) , bist e traet hylammonium hexabromo- and hexaiodo-dipalladate(11) and hexaiodo- diplatinate@) can be prepared,8 and we have also measured their crystal spectra. There has been con- siderable controversy about the assignment of the ligand- field transitions in square-planar platinum(I1) and palla- dium(I1) complexes, particularly with regard to which E.g., J.Lewis, Pure Appl. Chem., 1965, 10, 11. P. Day, A. F. Orchard, A. J. Thomson, and R. J. P. Williams, P. Day, A. F. Orchard, A. J. Thomson, and R. J. P. Williams, * D. S. Martin and G. A. Lenhardt, Inoug. Chem., 1964, 3, D. S. Martin, M. A. Tucker, and A. J. Iiassman, Inovg. J , Chem. Phys., 1965, 42, 1973. J . Chem. Phys., 1965, 43, 3763. 1368. Clzem., 1965, 4, 1682; 1966, 5, 1298. bands are spin-allowed and which spin-forbidden .s A source of some of the confusion has been the high spin- orbit coupling constant of platinum, which is expected to result in enhanced intensities for what would, in the absence of such coupling, be purely spin-forbidden transitions. The only spectrum which has been con- sidered in any detail, however, is that of PtC142-, and so an extension of the experimental information to the heavier halide complexes appeared desirable.EXPERIMENTAL Bistetraethylanimonium hexabromo- and hexaiodo-di- platinate(I1) and the corresponding dipalladates(11) were prepared according to the literature method.s Crystals suitable for transmission spectroscopy were obtained by slow recrystallisation from acetone, except in the case of hexaiododipalladate(II), for which slow evaporation of a solution in dimethylsulphoxide proved more effective. All four compounds form thin yellow-brown or dark brown plates but precise crystallographic information is avail- able only for [N(C,H,),],Pt,Br,. X-Ray powder photo- graphs of the hexabromo-dipalladate (11) and -diplatinate(rI) are almost identical, reinforcing the earlier suggestion 7 that they are isomorphous. The powder photographs of the hexaiodo-compounds have sufficient features in common with the hexabromo-analogues to suggest that they too have related structures, but no more detailed conclusions would be possible without a single-crystal X-ray examin- ation.This was not attempted. Polarised transmission spectra of the four compounds were measured on a single-crystal microspectrophotometer designed in our laboratory. A brief account of this in- strument has been given 2 and a more comprehensive des- cription of its design, construction and mode of operation has recently appeared.10 Briefly, the instrument is of the manual, single-beam type, in which the sample is mounted over a small hole (typically 50 p diameter) in a brass slide placed in the focal plane of a pair of Beck reflecting ob- jectives (magnification x 74).Light from either a 500 11- high pressure Xenon arc or a tungsten iodine lamp is monochromated by a Barr & Stroud double monochromater, polarised by a Glan-Thomson calcite prism, and, after passing through the microscope and sample, is detected either by a photomultiplier or a lead sulphide photo- conductive cell. 0. S. Mortensen, A d a Chem. S c a d . , 1965, 19, 1500. N. C. Stephenson, Acta Cryst., 1964, 17, 587. 8 C. M. Harris, S. E. Livingstone, and N. C . Stephenson, J . For a review, see H. B. Gray, Transition A4etal Chem., 1965, l o J . C. Barnes and A. J. Thomson, J . S C ~ . Tnstv., 1967, 44, 577. Chern. SOC., 1958, 3697. 1, 240.Inorg. Phys. Theor. 669 The procedure for obtaining a spectrum is as follows: a crystal is selected and examined under the polarising microscope to ensure that it is single and flawless.Then, still under the polarising microscope, it is mounted over the small hole and its extinction directions are marked on the side of the slide. The microscope in the spectrophoto- meter has a rotating stage so that the sample can be mounted with one of its extinction directions parallel to the axis of polarisation of the calcite prism. After a spectrum has been obtained with the incident electric vector along one of the crystal extinction directions, the polariser can be rotated through 90" and the spectrum remeasured along the other extinction direction. The crystal is then moved away from the hole and baselines are determined with the polariser in both positions, so that the spectra may be corrected for any polarisation inherent in the mono- chromater.Since all the crystals used in the present study were tri- clinic, their extinction directions bear no necessary relation- ship with the crystalline axes and, moreover, vary with wavelength. Strictly speaking, therefore, the correct procedure should be to determine the absorbance for a number of orientations of the polariser and hence calculate the absorptivity ellipsoid as a function of wavelength. An excellent example of this operation is found in the work of Stewart,ll on the crystal spectra of nucleotides, but in the present case, it is possible to make the assumption that the effective chromophore is square planar and hence that the absorbance with respect to the molecular axes can be defined by two parameters alone, ez = ey # e,.Then values of the absorbance along two orthogonal crystalline axes, held constant with wavelength, are sufficient to allow the calculation of absorbance with respect to the molecular axes if the crystal structure is known. For this purpose we have used the extinction directions defined in white light to orient the crystals as described above. The angles between the extinction directions and the known faces of a well-developed crystal may be found in a separate experiment under the polarising microscope. gives the equation of the least-square plane through the Pt,Br,2- group, referred to the triclinic crystalline axes. Optical goniometry on well developed crystals showed that the major face of the crystals, perpendicular to which all spectra were measured, was (1 00), hence Stephenson's least-squares equation was transformed to an orthogonal co-ordinate system, one axis of which was the crystallo- graphic c-axis while the new b*-axis remained in the crystallographic bc-plane.With the extinction directions on the (100) face defined with respect to the c-axis, the observed absorbances can then be transformed to absorb- ances within the molecular system of co-ordinates. To convert absorbance values to molar extinction coefficients, the thicknesses of the crystals were measured using a microscope with a micrometer eyepiece. The maximum error in reading the micrometer was f0.85 p and crystal thicknesses were in the range 10-50 p.No correction to the absorbance was made for reflection from the surfaces of the crystal. Solution spectra of the complexes were measured in acetone on a Unicam SP 700 and powder reflectance spectra on a Unicam SP 500 using samples finely ground with the appropriate potassium halide as diluent. Stephenson l1 R. F. Stewart and N. Davidson, J . Chern. Phys.. 1963, 39, 255. RESULTS AND DISCUSSION Figure 1 shows the polarised crystal spectrum of [N(C,H5),],Pt,Br,, both as measured along the two 3 .- '. I -- Freqwncy (kK) FIGURE 1 The polarised crystal spectrum of [N(C,H,),],Pt,Br,. The xy-spectrum is measured with the electric vector of the incident light along the crystal extinction direction which lies close to the molecular plane, the z-spectrum along the other extinction direction.The dotted lines represent absorption spectra calculated for the electric vector incident parallel and perpendicular to the crystallographic mean-square planes of the molecules 400i 1 c .Q c t W t Jv Frequency (kK) FIGURE 2 The polarised crystal spectrum of [N(C,H,),],Pt,I,. The xy and z labels have the same meaning as in Figure 1 extinction directions and resolved along the molecular axes. Since the anion makes an angle between 10-20" with the microscope axis the measured spectrum labelled xy contains a small amount of absorption polarised out of the plane of the anion and that labelled x contains some polarised in the plane. The dotted lines show the670 J. Chem. SOC. (A), 1968 effect of carrying out the resolution along the molecular axes.Most marked is the disappearance of the out-of- plane components of the bands at 16.3 and 234 k K . No such resolution was attempted for the other three Freqwncy ( k ~ ) FIGURE 3 The polarised crystal spectrum of [N(C,H,),],Pd,Br,. See Figure 1 for explanation of labels 0 25 20 I5 10 FIGURE 4 The polarised crystal spectrum of [N(C,H,),j,Pd,I,. See Figure 1 for explanation of labels compounds since detailed crystallographic information is not available; Figures 2 4 therefore simply report the spectra measured along the extinction directions. It is worth noting, however, that the in-plane and out-of- plane polarisations do not appear to be so well separated in the palladium as in the platinum compounds. In discussing the assignments of all these spectra and the question of metal-metal interaction it is helpful to consider spectra measured in solution as well as in the solid state, both for the dimers and those monomers for which data are available.We have therefore collected a selection of relevant data, both from the literature and from our own measurements, in Table 1. As a result TABLE 1 Solution and solid spectra of halogeno-palladates(r1) and Molar extinction coefficients are -platinates(Ir) (kK). given in parentheses PdBr,2- Pd2Br62- Powder Powder Solution reflect- Solution reflect- (CH,CN) a ance b (acetone) ,- ance c Crystal c 16.0 14.4 14.7 (410, xy) 20.0 19.0 (912) 19.1 19.4 (2630, X Y ; 24.4 (1700) 26.0 23.3 (5210) 24.0 23.7 (4940, X Y ; 30.1 (1 0,400) 30.4 (9340) 1080, 2) (20) 3300, 2 ) 40.5 (30,400) 36.5 35.5 PdzI,2- Powder PdI,2- Solution reflect- Crystal Solution (acetone) ance 14.3 14.4 (XY > Z ) (18.1) 18.5 (7090) 18.1 17.7 (XY > Z) 20.5 22.3 (10,000) 22.0 22.0 (XY > Z ) 24.9 29.4 (44,700) 28.7 PtBr,2- Pt2BreZ- Powder Solution reflect- Solution (1MBr-) b ance e (acetone) 16.7 (5) (16.5) (16.0) (31) 19-7 (15) 18-8 18.4 (71) 24.3 (100) (24.5) 23.4 (423) 28.2 (120) 26.3 30.4 (3900) 37.3 (7000) (32.5) Powder reflect- ance = Crystal c 16.5 16-3 (45, XY) 18-3 18.4 (63, xy; 20.5 (40, X Y ; 25, 3) 20, z ) 23.5 (23.5) (285, XY) 31.0 28 (400, X ~ Z ) PtzI,2- Powder PtI,2- Solution reflect- Solution (acetone) ance c Crystal c ( 16.2) (16.5) 16.0 (54, XY) (18.6) (18.2) (91) (18.5) 18.6 (70, X Y ; (22.5) (22.0) (795) (21.5) 23.0 (265, X,V; 40, z ) 70, z ) 25.7 27.5 (8140) 26.7 30.8 (31.5) (2180) 29.2 35.8 37.3 (1780) 35.0 0 C .M. Harris, S. E. Livingstone, and I. H. Reece, J. Chew. Soc., 1959, 1505. b H. B. Gray and C. J. Ballhausen, J . Ameu. CJzem. SOC., 1963, 85, 260. d C. K. Jorgen- sen, ‘ Absorption Spectra and Chemical Bonding,’ Pergamori, Oxford, 1962. e Ref. 2. c Present work. of polarised crystal spectroscopy 234-6 and Faraday effect measurements,12 the assignment of the ligand-field bands of PtC142- is now universally agreed to be as shown in Table 2, where the energies and polarisations of the bands of both PtC1,2- and PdC1,2- are also dis- played. These may then be taken as the point of de- parture for assigning the new spectra, with the addition of l2 D. S. Martin, J . G.Foss, M. E. McCarvilie, M. A. Tucker, and A. J. Iiassman, Inoyg. Chem., 1966, 5, 491.Inorg. Phys. Theor. 671 one further point : the charge-transfer bands. Gray and Ballhausen suggested that the bands in the solution spectra of PdCI,,- and PtC1,2- at 36.0 and 46.0 k K were l,4,, + lA2, ligand to metal transitions. We have shown that in a series of platinum salts the energy of this transition was sensitive to crystalline environment, and thus it can be identified with the peaks in the re- flectance spectra of I<,PtCl, and K,PdCl, at 42.5 and 36.5 k K . Polarised single-crystal reflection spectra l3 confirm that the band is out-of-plane polarised, as predicted for lA2,. The polarised crystal spectrum of Pt,Br62- bears a strong resemblance to that of PtC1,2-, and suggests that the bands may therefore be assigned by a straightforward comparison.Unfortunately, a t least formally, there are complications. If it is assumed that the dimeric com- plexes have D, symmetry, transitions from the lA,, ground state become electric-dipole allowed to excited states composed of antisymmetric combinations of d orbitals on each metal atom. However, inspection combinations of those d orbitals which in the local symmetry of each metal would be xy and z2 transform under the same irreducible representations of D,, (LZ~, and b2J. Hence, in contrast to the situation in MX,2- (D,,) , the one-electron vibronic transitions xy -+ x2 - y 2 and z2 + x2 - y2 should have similar polarisations. If the bands appearing only in xy polarisation at 16.3 and 23.5 kK in Pt,Br62- are to be related to the 3A2, and '4, (D,,)-transitions of PtC1,2-, seen in xy polarisation at 17.3 and 26.0 kK, then the inescapable conclusion is that even vibronically, the metal atoms are not coupled, and that the selection rules remain as for a single isolated square planar ion.This conclusion becomes comprehensible in the light of a recent far infrared spectral study of Pt2XG2- ions,15 in which it was demonstrated that the difference between bridging and terminal metal-halogen stretching force constants becomes smaller with increasing halogen atomic number. Both the solution and crystal extinction coefficients TABLE 2 and molar extinction coefficients are in parentheses PtC1,2- PdC1,2- Solution and solid spectra of tetrachloro-palladate(I1) and -platinate(II), with assignments.Band energies are in l w 7 Solution a Reflectance a Crystal GGZZ--Keflectance Crystal 7 h r 3'42g ............... 17.7 (3) (17.5) (15.3) (5, X Y ) 3E, ............... 21.0 (15) (20-4) 20.3 (17, X Y ; 20, Z ) 16.7 17.5 (19, xy; 7, z ) 3B,, ............ 23-8 (25, XY, 15, Z ) lA2, ............... 25.5 (59) (27.0) (26.0) (45, XY) 22.1 (200) 21.5 20.0 (67, xy) ' E , ............... 30.2 (64) 29.0 28-8 (57, X Y ; 70, Z ) 25.0 (250) 23.3 22.8 (125, X Y ; 80, Z ) lBIg ............... 37.9 (250) (36.5) 30.2 (1200) (31.5) 29.5 (67, X Y ) lA2* '............... 46.0 (9580) 42.5 36.0 (12,000) 36-5 a J. Chstt, G. A. Gamlen, and L. E. Orgel, J . Chem. SOC., 1958, 486. b Ref. 2. C. M. Harris, S. E. Livingstone, and I. H. Reece, J .Clzem. SOC., 1959, 1505. d H. B. Gray and C. J. Ballhausen, J. Amer. Chem. SOC., 1963, 85, 260. of the matrix elements of the electric-dipole moment operator between the ground state and such excited states reveals that the only non-zero terms are of the type ((9 - Y~)~IYI(x~)~), etc., where A and B are the two metal atoms. Although these terms are not re- quired to be zero, it is evident from the similarity be- tween the polarised spectra of PtCl,,- and Pt2Br6,- that they must be vanishingly small. Thus we have stated, in effect, that from the spectroscopic point of view there is negligible interaction between the metal atoms. ,4 second complication concerns the vibronic mech- anism of intensity borrowing between the ligand-field and charge-transfer states. If a vibronic analysis is carried out under the assumption of full D,, .symmetry, it is concluded that all ligand field transitions gain intensity in all three directions of polarisation, a pre- diction clearly incompatible with the experimental evi- dence of Figure l.It is not necessary to reproduce the argument in full, or to enumerate the forms taken by the symmetry-adapted molecular orbitals, normal vibr- ations l4 or transition-dipole matrix elements; it is sufficient to see that both symmetric and antisymmetric 13 B. G. Ahex, hl. E. Ross, and M. W. Hedgcock, J . Chem. Phys., 1967, 46, 1090. of the dimeric anions are generally more than twice those of the corresponding monomers, so there might be an additional contribution to the intensity of the ligand field transitions from a small static distortion.Prob- ably as a result of the lower negative charges on the two bridging bromide ions, the angles about the platinum atoms in Pt,Br62- deviate somewhat from go", so that the true site symmetry of the platinum atoms is close to C2L. Transitions from xy, yz, and z2 to x2 -y2 then become electric-dipole allowed, with polarisations in, out of, and in the molecular plane, respectively. The effect of any intensity gained in this way would simply be to distort the polarisation ratios arrived at vibronically, but since we have a t present no means other than group theory for deciding what these ratios should be, no further discussion of the question is pos- sible. Such a mechanism would not introduce out-of- plane intensity into the xy + x2 - y2 transition.If we are correct in rejecting the conclusions of the vibronic analysis for the full dimer, and concentrate only on the local site symmetry of the metal ions, the lower excited states of Pt,Br62- are easily assigned, as l4 R. P. Bell and H. C. Longuet-Higgins, Pmc. Roy. SOC., 1945, -4, 183, 357. l5 H. Basch and H. B. Gray, Inovg. Chenz., 1967, 6, 365.672 J. Chem. SOC. (A), 1968 follows (all transitions are from lA1, in D4,J: 16.3, 3A,; 18.4, 3E,; 20.8, 3B1,; and (23.5), 1A2g. This agrees with the conclusions arrived at by Basch and Gray,lG who argued from an SCCC-MO calculation and the available powder reflectance data for K2PtBr4. The next band, occurring at 30.4 in the solution spectrum and 31.0 kK in the powder reflectance spectrum, is too intense to be observed in the polarised crystal spectrum, though it is clearly of mixed polarisation.In the solu- tion spectrum of PtBr42- there is a band at 28.2 kK with an extinction coefficient of 120 which clearly re- lates, both in energy and intensity, to the band of PtC142- whose Faraday effect defines it as lE,. However, the Pt2Br62- band in this region is 200 times more intense than that of PtBr42-, so we are inclined to assign it, not as lEg, but as the first ligand-to-metal electron transfer band, lBlu in D2h. Hence the most striking consequence of going from PtBr,2- to Pt2Br62- is that this band has fallen in energy by about 7 k K . 16.0, 3A2g; 18.6, 3Eg; and 23.0, 1A2g, though our formal classification of the excited states as singlets or triplets has little physical significance for atoms as heavy as platinum and iodine.Indeed, the intensities of the ' spin-for- bidden ' ligand-field transitions in Pt I 2- are actually greater than some of the ' spin-allow:;' transitions of PtC142-. Martin, Tucker, and Kassman 5 have used the strong-field matrix elements of the d* configuration in a square-planar ligand field, including electron-electron repulsion and spin-orbit coupling 17 to assign the spectrum of PtC142-, but even at 1 6 " ~ only a fraction of the pre- dicted number of transitions is observed. This fact, together with the strong qualitative similarity between the lower energy bands of the chloro-, bromo-, and iodo- platinates is good evidence that, whilst spin-orbit coupl- ing is an important factor determining the intensities of the formally spin-forbidden transitions, spin-orbit components of the various excited configurations are not sufficiently separated in energy to present detectable spectral features.In PdC1,2- the ligand field transitions are already significantly more intense than in PtC142-, a fact attri- buted to the lower energy of the electron transfer states in the palladium compound. The solution spectra of PdBr42- and Pd2Br62-, as well as the crystal spectrum of the dimer, have a well resolved band near 24 kK which, if it were not for the fact that its extinction coefficient lies between 2000 and 4000, would be assigned at once as lEg by analogy with PdC142-. The lower energy bands in the polarised crystal spectrum of Pd2Br62-, though distributed in energy and polarisation in a manner very similar to PdC1,2-, are also extremely intense. The most probable explanation is that the electron transfer states from which the ligand-field transitions gain their intensity are now sufficiently low in energy that vibronic l6 D.M. Adams, P. J. Chandler, and R. G. Churchill, J . Chew. SOC. ( A ) , 1967, 1272. The assignment for Pt2162- is similar: mixing has all but cancelled the electric dipole selection rule. Thus if an asymmetric vibrational co-ordinate Q mixes a u-function into the g-functions, the extent of mixing is given by first-order perturbation theory : where V is the crystal field potential. The oscillator strength of a transition from #ls to $28 then becomes 18 Going from chloride to bromide to iodide, a is expected to decrease slightly as the potential curve of the molecule becomes flatter, but the overwhelming effect is the prox- imity of u-excited states to g-excited states. If the energy of a ligand field state remained the same at 20 kK, while a charge transfer state fell from 35 to 25 kK, equation (2), predicts that the oscillator strength of the former would rise by a factor of 6.5.N7e therefore believe that the lower bands of Pd2BrG2- and Pd21G2- can be classified as ligand-field transitions, despite their highintensity. Following the scheme derived for PdC1,2-, we assign the bands of Pd2Br62- and Pd2162- at 14.9 and 14.3 kK as 3E,, those at 19.1 and 17.7 kK as 1A2g. Some doubt must remain as to whether the bands a t 22.8 and 22.0 krc in the bromide and iodide represent lEg, or the first electron-transfer bands. CONCLUSION This work shows that the well established assign- ments of the ligand-field transitions of PtC142- and PdC1,2- are readily extended to square-planar complexes of the heavier halogens. With the spectral resolution attainable at room temperature, no complications from increasing spin-orbit coupling have become apparent, except for the increased intensity of the spin-forbidden transitions. In the bronio- and iodo-palladates, vibr- onic coupling with neighbouring electron-transfer states increases the intensities of the ligand-field transitions very noticeably. The absence of axial stacking of the planar anions in these crystals does not lead to any noticeable spectroscopic effects when comparisons are made with K,PtC14 and K2PdC14, but the extent to which the polarisations of the bands are determined by the vibronic properties and local symmetry of the individual metal ion co-ordination sites is, perhaps, surprising. No evidence has been found for any electronic inter- action between the metal ions within each dimer. [7/1040 Received, August 9th, 19671 17 R. F. Fenske, D. S. Martin, and I<. Ruedenberg, Inorg. l8 C . J. Ballhausen, ' Introduction t o Ligand Field Theory,' Chem., 1962, 1, 441. McGraw-Hill, New York, 1963, p. 186.

ISSN:0022-4944    

DOI:10.1039/J19680000668

出版商:RSC    

年代:1968

数据来源: RSC